CN100449764C - Photodetector - Google Patents

Abstract

A photodetector comprising a semiconductor substrate with a photosensitive unit (1a, 1b, 1c). Each photosensitive cell is provided with a filter layer 20 that transmits light within a predetermined wavelength range of the photosensitive cell, and a photoelectric conversion unit 17 operable to generate signal charges according to the intensity of light transmitted through the filter layer 20 . The thickness (ta, tb, tc) of the filter layer 20 corresponds to the predetermined wavelength range of the respective photosensitive unit. With this structure, it is possible to provide a cost-effective photodetector that does not require material management of pigments and dyes of different colors when manufacturing a color filter.

Description

Photodetector

technical field

The present invention relates to a photodetector such as an image sensor, and more particularly to a color filter arranged in a photosensitive unit of the photodetector.

Background technique

A commonly used photodetector includes a semiconductor substrate with a plurality of photosensitive units, each photosensitive unit is provided with a color filter and a photodiode. There are red color filters (R), green color filters (G), and blue color filters (B) in the photodetector, and a color filter of one of the three colors is assigned to each photosensitive unit. The photodiode generates a signal charge in response to the intensity of light passing through the color filter. Thus, signal charges are generated in each photosensitive cell in response to light of a particular color in the incident light. Imaging data of an image is generated by collecting signal charges of three colors generated in a photosensitive cell (see, for example, Japanese Laid-Open Patent Application No. H05-006986).

Conventional color filters are generally produced by mixing acrylic resins with pigments or dyes of respective colors (see, for example, Japanese Laid-Open Patent Application No. H07-311310).

However, manufacturing a conventional color filter requires cumbersome material management for different colors of pigments or dyes during the manufacturing process, which prevents cost reduction of photodetectors.

Contents of the invention

In view of the above problems, it is an object of the present invention to provide a cost-effective photodetector that does not require material management of pigments or dyes during the manufacturing process.

In order to achieve the above object, the photodetector of the present invention includes a semiconductor substrate having a plurality of photosensitive units, each photosensitive unit comprising: a filter layer for transmitting light within a predetermined wavelength range of the photosensitive unit; and a photoelectric conversion unit operable for Signal charges are generated according to the intensity of transmitted light, wherein the thickness of each filter layer corresponds to the predetermined wavelength range of each photosensitive unit, and each filter layer in all photosensitive units is made of only one material.

With the above structure, the wavelength range of light transmitted through the filter layers is determined on the basis of the thickness of each filter layer corresponding to a predetermined color. Therefore, material management of pigments or dyes of different colors is not necessary in the manufacturing process, so that the production cost can be reduced.

The above-mentioned photodetector may also be of such a structure, wherein the filter layer is made of only one material whose transmittance is lower than the transmittance when the light wavelength is equal to or greater than the cut-off wavelength, and the cut-off The wavelength is determined by the thickness of the material.

With the above structure, the filter layer can have the function of a color filter that mainly transmits light with a wavelength greater than the cutoff wavelength.

The above-mentioned photodetector may also be a structure in which the filter layer is made of a material whose transmittance is reduced by absorbing light.

In order to reduce light transmission, there are two possible approaches. One is to make the filter layer absorb light, and the other is to make the filter layer reflect light. In the latter method, the light reflected by the filter layer may be scattered in the photodetector, so the occurrence of the flare phenomenon is higher than in the case of the former method. Glare is the phenomenon where a correct signal overlaps with an incorrect signal of light reflected from adjacent photosensitive cells. With the present invention, the occurrence rate of the glare phenomenon is reduced by using the former method.

The aforementioned photodetector may also have a structure in which the filter layer is made of one of polysilicon, amorphous silicon, and silicon.

With the above structure, the filter layer can be manufactured using semiconductor manufacturing technology. For example, a polysilicon layer is formed on the photosensitive unit using CVD technology, and then dry-etched so that each layer has a predetermined thickness for each photosensitive unit.

Therefore, a color filter manufacturing process dealing with acrylic resin is unnecessary. Also, it is possible to utilize manufacturing equipment for semiconductor manufacturing and simplify the manufacturing process.

The aforementioned photodetector may also have this structure, wherein the substrate is made of silicon, the photoelectric conversion unit is formed by doping the substrate with N-type impurities, and the filter layer is formed by doping the photoelectric conversion unit with P-type impurities.

With the above structure, the filter layer can be manufactured using semiconductor manufacturing technology. For example, the filter layer can have the required thickness of each photosensitive unit by adjusting the acceleration energy of the P-type impurities.

Therefore, a color filter manufacturing process dealing with acrylic resin is unnecessary. In this way, manufacturing equipment for semiconductor manufacturing can be utilized and the manufacturing process can be simplified.

The above-mentioned photodetector may also have a structure in which the filter layer is made of a material whose cutoff wavelength becomes larger as the material becomes thicker.

With the above structure, it is possible to adjust the transmitted light in a desired wavelength range by determining the thickness.

The above-mentioned photodetector can also have this structure, wherein the thickness of each filter layer is one of the first thickness, the second thickness and the third thickness, and the cut-off wavelength of the first thickness is in the red wavelength range and the green wavelength range Between, the cutoff wavelength of the second thickness is between the green wavelength range and the blue wavelength range, and the cutoff wavelength of the third thickness is between the blue wavelength range and the ultraviolet wavelength range.

According to the above structure, each filter layer provided in each photosensitive unit can transmit light in one of the following wavelength ranges according to the thickness of the filter layer.

First Thickness: Red Light Wavelength Range

Second thickness: red and green wavelength range

Third Thickness: Red, Green, and Blue Wavelength Ranges

Specifically, signal charges corresponding to light in the red wavelength range are generated in the first photosensitive unit provided with a filter layer having a first thickness. Similarly, signal charges corresponding to light in the red and green wavelength ranges are generated in a second photosensitive unit provided with a filter layer having a second thickness, and corresponding to red, green Signal charges for light in the wavelength range of and blue light are generated in a third photosensitive unit provided with a filter layer having a third thickness.

Thus, a red signal is derived from the signal charge generated in the first photosensitive cell. And, a green signal is derived from a difference between signal charges generated in the first and second photosensitive cells. Likewise, the blue signal is derived from the difference between the signal charges generated in the second and third photosensitive cells.

The photodetector above can also be of this structure, wherein the photosensitive unit further includes an anti-reflection layer made of a material with a lower refractive index than the filter layer, and the anti-reflection layer is disposed on the main surface of the filter layer facing the light source.

With the above structure, the anti-reflection layer is disposed between the filter layer and the gas, and the reflectance of incident light can be reduced. Therefore, the sensitivity of the photodetector can be improved.

The above-mentioned photodetector may also have this structure, wherein the filter layer is made of one of polysilicon, amorphous silicon, and silicon, and the antireflection layer is made of one of silicon nitride, silicon dioxide, and silicon oxynitride.

With the above structure, the filter layer and the anti-reflection layer can be manufactured using semiconductor manufacturing technology. Therefore, it is possible to utilize manufacturing equipment for semiconductor manufacturing and to simplify the manufacturing process.

The above-mentioned photodetector may also be of such a structure, wherein the photosensitive unit further includes a light-shielding forming layer having a light-shielding layer and a small hole at a portion corresponding to the photoelectric conversion unit, the light-shielding layer blocks light other than passing through the small hole, and filters light. The layer is located between the light-shielding forming layer and the photoelectric conversion unit.

With the above structure, it is possible to prevent light scattered by adjacent photosensitive cells from entering the photoelectric conversion cells. Thus, in each photosensitive cell, noise caused by scattered light from neighboring photosensitive cells is reduced.

The aforementioned photodetector may also have this structure, wherein the photosensitive unit further includes a silicon dioxide layer with a thickness in the range of 1 nm to 150 nm, disposed between the filter layer and the photoelectric conversion unit.

With the above structure, the filter layer and the photoelectric conversion unit can be insulated. In this way, the signal charge of the photoelectric conversion unit can be prevented from leaking to the filter layer.

In general, incident light is reflected with a certain reflectance at the boundary surface between the silicon dioxide layer and the photoelectric conversion unit. However, it is known that the reflectance can be reduced by setting the thickness of the silicon dioxide layer within the range of 1 nm to 150 nm. Therefore, the decrease in the sensitivity of the photosensitive cell can be suppressed by interposing the silicon dioxide layer.

The photodetector described above may also be of a structure wherein the photosensitive unit further includes: a gate operable to generate a gate potential in a gate region between the photoelectric conversion unit and the transfer target region when the signal charge is not transferred , the gate potential is lower than the potential at the photoelectric conversion unit; and a potential barrier is disposed between the filter layer and the photoelectric conversion unit, and a potential barrier potential lower than the gate potential is generated at the potential barrier.

With the above structure, during the non-transfer time, the potentials at the gate region, the photoelectric conversion unit, and the potential barrier are low, high, and low, respectively. In this way, electrons generated in the photoelectric conversion unit are accumulated in the photoelectric conversion unit having a higher potential. Also with this structure, accumulated electrons can be prevented from leaking over the potential barrier to the filter layer.

Further, in order to achieve the above object, the photodetector according to the present invention includes a semiconductor substrate having a plurality of photosensitive units, each photosensitive unit includes: a photoelectric conversion unit operable to generate signal charges according to light intensity, some of which photosensitive units Each is provided with a filter layer through which light is transmitted to the photoelectric conversion unit, the thickness of each filter layer corresponds to the predetermined wavelength range of each photosensitive unit, and each filter in all the photosensitive units The layers are all made of only one material, while the remaining photosensitive cells are not provided with filter layers.

With the above structure, similar to the previously described photodetector, material management of pigments or dyes of different colors is unnecessary in the manufacturing process, so that the production cost can be reduced.

The above-mentioned photodetector may also have a structure in which the filter layer is made of a material whose transmittance to light having a wavelength smaller than the cutoff wavelength is lower than the transmittance to light having a wavelength equal to or greater than the cutoff wavelength , the cutoff wavelength is determined by the thickness of the material.

With the above structure, the filter layer can have the function of a color filter that mainly transmits light with a wavelength greater than the cutoff wavelength.

The above-mentioned photodetector can also have this structure, wherein the thickness of each filter layer is one of the first thickness and the second thickness, the cut-off wavelength of the first thickness is between the red light wavelength range and the green light wavelength range, and the second thickness is between the red light wavelength range and the green light wavelength range. The cut-off wavelength of the second thickness is between the green light wavelength range and the blue light wavelength range.

This structure can be applied to the case where the photodetector receives light from which ultraviolet rays have been removed. Usually, in an ordinary imaging system, after passing through an optical low-pass filter and an infrared cut filter, the incident light collected by a lens reaches a photodetector. In such an imaging system, if a UV cut filter is also provided, the UV cut filter serves to transmit light of a wavelength range whose wavelength is greater than the boundary between the blue wavelength range and the UV wavelength range. Therefore, no filter layers are required in photosensitive cells that generate signal charges based on light in this wavelength range.

Thus, the manufacturing process of the filter layers can be shortened because it is sufficient to form filter layers for only two colors for the photodetector without forming filter layers for three colors.

The signal processing device according to the present invention obtains a color signal based on a first source signal and a second source signal among a plurality of source signals produced by the photodetector as claimed in claim 1, the first source signal corresponding to the first wavelength range The light intensity in the second source signal corresponds to the light intensity in the second wavelength range including the first wavelength range, and the color signal corresponds to the light intensity in the second wavelength range excluding the first wavelength range. The signal processing device includes: a holding unit which holds weighting coefficients respectively corresponding to the first source signal and the second source signal; Differences are derived between the results to obtain color signals.

With the above structure, the signal processing means can obtain a color signal corresponding to the light intensity in the second wavelength range excluding the first wavelength range. The above-described signal processing means is used when the photodetector outputs the first source signal and the second source signal as described above.

According to the signal processing device of the present invention, the first source signal, the second source signal and the third source signal among the plurality of source signals generated by the photodetector according to claim 1 are obtained corresponding to the light in the wavelength range of red light. A strong red signal, a green signal corresponding to light intensity in the green wavelength range, and a blue signal corresponding to light intensity in the blue wavelength range, the first source signal corresponding to a wavelength range greater than the red wavelength range and the green wavelength range The light intensity of the wavelength between, the second source signal corresponds to the light intensity of the wavelength range greater than the wavelength range between the green light wavelength range and the blue light wavelength range, and the third source signal corresponds to the wavelength range greater than between the blue light wavelength range and the ultraviolet wavelength range The light intensity between wavelengths, the signal processing device includes: a holding unit, which holds a conversion matrix for converting a group of first source signals, second source signals and third source signals into a group of red signals, green signals and blue signals; and a computation unit operable to apply a transformation matrix to the set of first, second and third source signals.

Each source signal corresponds to the intensity of light in the wavelength range described below.

First source signal: red light wavelength range

Second source signal: red and green wavelength range

Third source signal: red, green and blue wavelength ranges

According to the above structure, the signal processing device can obtain a group of red signals, green signals and blue signals according to the group of first source signals, second source signals and third source signals, as follows:

Red signal: red light wavelength range

Green signal: Green light wavelength range

Blue signal: blue light wavelength range

A method of manufacturing a photodetector according to the present invention, which includes forming a semiconductor substrate with a plurality of photosensitive units, each photosensitive unit is provided with a filter layer that transmits light within a predetermined wavelength range of the photosensitive unit, and photoelectric conversion A unit operable to generate a signal charge according to the intensity of transmitted light, the method comprising: forming a photoelectric conversion unit on each photosensitive unit; forming a layer of the same thickness on each photosensitive unit, the layer being made of a material , the wavelength range of the transmitted light of the material is determined based on the thickness; and forming each filter layer made of only one material in all photosensitive units by etching said layer so that the resulting layer has predetermined thickness.

With the above structure, the wavelength range of light transmitted through the filter layer is determined based on the thickness of each layer corresponding to a predetermined color. Therefore, material management of pigments or dyes of different colors is not necessary in the manufacturing process, so that the production cost can be reduced.

Further, if polysilicon or amorphous silicon is selected as the material constituting the filter layer, the filter layer can be manufactured using semiconductor manufacturing techniques. Therefore, it is not necessary to manage the color filter manufacturing process of acrylic resin. As a result, manufacturing equipment for semiconductor manufacturing can be utilized and the manufacturing process can be simplified.

A method of manufacturing a photodetector according to the present invention, wherein the photodetector includes a semiconductor substrate having a plurality of photosensitive units, each photosensitive unit is provided with a filter layer that transmits light within a predetermined wavelength range of the photosensitive unit, and a photoelectric a conversion unit operable to generate signal charges according to the intensity of transmitted light, the method comprising: forming a photoelectric conversion unit by doping an N-type impurity in each photosensitive unit; and forming a photoelectric conversion unit by doping a P-type impurity in the photoelectric conversion unit Each filter layer made of only one material is formed in all photosensitive cells so that the impurity portion has a predetermined thickness according to each photosensitive cell.

With the above structure, the wavelength range of light transmitted through the filter layer is determined based on the thickness of each layer corresponding to a predetermined color. Therefore, material management of pigments or dyes of different colors is unnecessary in the manufacturing process, so that the production cost can be reduced.

Further, the filter layer can be manufactured using semiconductor manufacturing technology. Therefore, it is unnecessary to manage the color filter manufacturing process of acrylic resin. In this way, manufacturing equipment for semiconductor manufacturing can be utilized and the manufacturing process can be simplified.

Description of drawings

These and other objects, advantages and features of the invention will become apparent from the following description taken in conjunction with the accompanying drawings, which illustrate specific embodiments of the invention.

In the attached picture:

Figure 1 illustrates the structure of an imaging system according to the present invention;

Figure 2 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a first embodiment;

3 is a graph showing the light transmittance of a polysilicon layer and an infrared cut filter;

FIG. 4 is a block diagram illustrating an internal structure of a signal processing circuit;

5A-5B show the matrix and inverse matrix held in the matrix holding unit 61;

6A-6H have illustrated an example of the method for manufacturing optical filter layer 20;

Figure 7 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a second embodiment;

Figure 8 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a third embodiment;

9A and 9B show the potentials of the filter layer 20, potential barrier 23, photoelectric conversion unit 17 and transfer transistor 24 in the photosensitive unit;

Figure 10 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a fourth embodiment;

11A-11C show the relationship between the potential and the distance from the boundary surface between the insulating layer 13 and the photodiode layer 12 in the photosensitive unit;

12A-12G illustrate another example of the method for making the filter layer 20; and

Fig. 13 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a modified embodiment.

Detailed ways

first embodiment

[summary]

In the first embodiment, the filter layers in all pixel units are only made of polysilicon. As a characteristic of polysilicon, the wavelength range of light transmitted through a polysilicon layer varies according to the thickness of the layer. Therefore, by setting the thickness of the polysilicon layer different from each other for each photosensitive unit, the polysilicon filter layer can be used as a color filter.

Since the wavelength range of light transmitted through the polysilicon filter layer is determined based on the thickness of the corresponding filter layer rather than each color of pigment or dye, there is no need for different color pigments or dyes in the process of manufacturing the filter layer. material control. Therefore, production costs can be reduced.

In addition, the filter layer can be manufactured using semiconductor manufacturing technology, and it is not necessary to manage the color filter manufacturing process of acrylic resin. As a result, manufacturing equipment for semiconductor manufacturing can be utilized and the manufacturing process can be simplified.

[structure]

Fig. 1 illustrates the structure of an imaging system according to the present invention.

The imaging system includes a photodetector 1 , a driving circuit 2 , a vertical scanning circuit 3 , a horizontal scanning circuit 4 , an analog front end 5 , a signal processing circuit 6 , a working memory 7 , a recording memory 8 , and a control unit 9 .

The photodetector 1 is a so-called MOS image sensor consisting of photosensitive cells (1a, 1b, 1c) formed on a semiconductor substrate. Signal charges are generated in each photosensitive unit according to the intensity of light within a predetermined wavelength range of the corresponding photosensitive unit. The photosensitive cells of Figure 1 are labeled R, RG, and RGB, respectively. "R" indicates that the unit is equipped with color filters that transmit light in the red wavelength range, "RG" indicates that the unit is equipped with color filters that transmit light in the red and green wavelength ranges, and "RGB ” indicates that the unit is provided with color filters that transmit light in the red wavelength range, the green wavelength range and the blue wavelength range. Specifically, signal charges corresponding to light in the wavelength range of red light are generated in the photosensitive cell 1a. Similarly, signal charges corresponding to light in the red and green wavelength ranges are generated in the photosensitive unit 1b, and signal charges corresponding to light in the red, green and blue wavelength ranges are generated in the photosensitive unit 1c. Also, as shown in FIG. 1, the color filters are basically arranged based on the Bayer pattern diagram. In the case of using the Bayer pattern diagram, color filters for three different colors are assigned to four pixel units. However, the arrangement of color filters is not limited to the Bayer pattern, and color filters for four different colors may be assigned to four pixel units, respectively.

In this specification, the wavelength range of blue light is defined as less than 490nm but not less than 400nm, the wavelength range of green light is defined as less than 580nm but not less than 490nm, and the wavelength range of red light is defined as less than 700nm but not less than 580nm. Further, the wavelength range of less than 400nm is defined as the ultraviolet wavelength range, and the wavelength range of 700nm or greater is defined as the infrared wavelength range.

The driving circuit 2 drives the vertical scanning circuit 3 and the horizontal scanning circuit 4 according to a trigger signal from the control unit 9 .

The vertical scanning circuit 3 activates the photosensitive cells sequentially and row by row based on the driving signal from the driving circuit 2 . Then, the vertical scanning circuit 3 simultaneously transmits the signal charges in a row of activated photosensitive cells to the horizontal scanning circuit 4 .

The horizontal scanning circuit 4 and the vertical scanning circuit 3 operate simultaneously based on the driving signal from the driving circuit 2 . For each column, the horizontal scanning circuit 4 sequentially outputs the signal charges transmitted by one row to the analog front end 5 .

Using the driving circuit 2 , the vertical scanning circuit 3 , and the horizontal scanning circuit 4 , the signal charges in the photosensitive cells arranged in the matrix are converted into signal voltages, and then the signal voltages are continuously output to the analog front end 5 .

The analog front end 5 samples and amplifies the signal voltage, converts the analog signal into a digital signal through A/D conversion, and then outputs the digital signal.

The signal processing circuit 6 is a digital signal processor (DSP). The signal processing circuit 6 converts the digital signal from the analog front end 5 into a red signal, a green signal, and a blue signal, and generates imaging data.

The working memory 7 is used when the signal processing circuit 6 converts digital signals corresponding to photosensitive cells into color signals of respective colors. A specific example of the work memory 7 is SDRAM.

The recording memory 8 records imaging data generated by the signal processing circuit 6 . A specific example of the recording memory 8 is SDRAM.

The control unit 9 controls the drive circuit 2 and the signal processing circuit 6 . For example, the control unit 9 outputs a trigger signal to the driving circuit 2 in response to the user pressing the shutter button.

The structure of the photosensitive cells (1a, 1b, 1c) in the photodetector 1 is explained in detail below.

Fig. 2 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a first embodiment.

Each photosensitive unit has a multi-layer structure disposed on a silicon substrate 11 doped with N-type impurities. These layers provided on the silicon substrate 11 are as follows.

The photodiode layer 12 includes a P-type well 16 formed by implanting P-type impurities into the substrate 11 , and an N-type region photoelectric conversion unit 17 formed by implanting N-type impurities into the P-type well 16 .

The insulating layer 13 is a silicon dioxide layer, and its purpose is to insulate the photodiode layer 12 and the light-shielding forming layer 14 .

The light-shielding forming layer 14 includes a light-shielding layer 18 composed of a metal such as aluminum and tungsten. After masking the portion corresponding to the small hole 19 in the light-shielding forming layer 14 with a mask, the light-shielding layer 18 is formed by metal deposition or sputtering. Furthermore, the light-shielding forming layer 14 may include, in addition to the light-shielding layer 18 , wires from the vertical scanning circuit 3 and/or to the horizontal scanning circuit 4 through which signal charges are transferred.

The filter forming layer 15 includes a filter layer 20 made of polysilicon.

The incident light 22 entering from the upper part of the photosensitive unit is collected by the microlens 21 , passes through the filter layer 20 and the small hole 19 and reaches the photoelectric conversion unit 17 . Note that the incident light 22 in the illustration has passed through an infrared cut filter (not shown in the drawings) above the photosensitive unit, so light in the infrared wavelength range has been cut off.

The thickness (ta, tb, tc) of the filter layer 20 is determined according to each photosensitive unit.

In general, polysilicon has a property that its cutoff wavelength is determined according to its thickness, and a polysilicon layer of a certain thickness transmits light of the cutoff wavelength and longer wavelengths while substantially cutting off light of a shorter wavelength than the cutoff wavelength. Here the cut-off wavelength refers to the wavelength at which the light transmittance begins to fall below a certain ratio of the peak transmittance. Strictly speaking, polysilicon also transmits a small amount of light with a wavelength smaller than the cutoff wavelength. However, such an amount of light is as small as zero and does not contribute much in forming signal charges. Therefore, for simplicity of description, light having a wavelength smaller than the cutoff wavelength is described as being cutoff in this specification.

In the case of polysilicon, as the thickness increases, the cutoff wavelength becomes longer. As the thickness of polysilicon becomes thinner, the cutoff wavelength becomes shorter.

Specifically, by adjusting the thickness of the filter layer 20 in each photosensitive unit, the photosensitive units can be divided between the wavelength range of red light and green light, between the wavelength range of green light and blue light, and between the wavelength range of blue light and ultraviolet light. (1a, 1b, 1c) Set cutoff wavelengths (λa, λb, λc).

When the cut-off wavelength is set in this way, the signal charges in the photosensitive unit 1a are generated based on light in the red wavelength range, the signal charges in the photosensitive unit 1b are generated based on light in the red and green wavelength ranges, and the photosensitive unit 1c The signal charge in is based on light generation in the red, green and blue wavelength ranges.

Specifically, the cut-off wavelengths (λa, λb, λc) of the filter layer 20 in the photosensitive unit (1a, 1b, 1c) are set as follows:

λa: 580nm

λb: 490nm

λc: 400nm

The thicknesses ta, tb and tc of the filter layer 20 corresponding to the listed cutoff wavelengths are as follows:

ta: 0.5 μm

tb: 0.3μm

tc: 0.1 μm

Thus, the wavelengths of light that can be transmitted through the filter layer 20 of the photosensitive unit are as follows:

Photosensitive unit 1a:

580nm or longer (red wavelength range)

Photosensitive unit 1b:

490nm or longer (red and green wavelength range)

Photosensitive unit 1c:

400nm or longer (red, green and blue wavelength range)

The above description is shown in FIG. 3 .

FIG. 3 is a graph showing light transmittance of a polysilicon layer and an infrared cut filter.

Curve 20a in the graph shows the transmission characteristics of polycrystalline silicon having a thickness of 0.5 μm through which light having a wavelength of 580 nm or longer is transmitted.

Curve 20b in the graph shows the transmission characteristics of polysilicon at a thickness of 0.3 μm through which light having a wavelength of 490 nm or longer is transmitted.

Curve 20c in the graph shows the transmission characteristics of polysilicon at a thickness of 0.1 μm through which light having a wavelength of 400 nm or longer is transmitted.

In the first embodiment, the thickness of the filter layer is set to be ta=0.5 μm, tb=0.3 μm, and tc=0.1 μm. However, actual values of ta, tb, and tc may be set within the ranges of 0.01 μm to 2.0 μm, 0.01 μm to 1.5 μm, and 0 μm to 1.0 μm, respectively.

The advantage of making these thicknesses thinner is that the sensitivity increases, the disadvantage is that the color separation becomes difficult because the difference between the thicknesses ta, tb and tc becomes smaller, and thus the difference in the transmission spectrum becomes smaller. On the other hand, although a greater difference in spectrum reduces sensitivity, the advantage is that the separation of colors is easier.

Curve 20d in the graph shows the transmission characteristics of an infrared cut filter with which light having a wavelength equal to or greater than 700 nm is cut.

Utilizing the above-mentioned transmission characteristics, among the incident light 22 to the photosensitive unit 1 a , only the light in the red wavelength range passes through the filter layer 20 and reaches the photoelectric conversion unit 17 . Similarly, of the incident light 22 to the photosensitive unit 1 b , only light in the wavelength range of red light and green light reaches the photoelectric conversion unit 17 . Also, of the incident light 22 to the photosensitive unit 1c, only light in the wavelength ranges of red light, green light and blue light reaches the photoelectric conversion unit 17.

The incident light 22 that has been transmitted through the filter layer 20 further passes through the small hole 19 .

The light shielding layer 18 is disposed above the photoelectric conversion unit 17 to prevent light scattered in adjacent photosensitive units from entering the photoelectric conversion unit 17 . The small hole 19 is positioned directly on the photoelectric conversion unit 17 . With the above structure, incident light substantially perpendicular to the substrate 11 reaches the photoelectric conversion unit 17, and light at an angle to the substrate is blocked.

The photoelectric conversion unit 17 is a photodiode formed by a pn junction having a p-type well 16 , and generates signal charges according to the intensity of light reaching the photoelectric conversion unit 17 through the filter layer 20 and the pinhole 19 . The mechanism of photoelectric conversion is as follows.

The photoelectric conversion unit 17 has a depletion region in which the carrier electrons are depleted after being combined with the carrier holes in the P-type well. With the above structure, the potential at the photoelectric conversion unit 17 increases relative to the potential at the P-type well 16, and therefore, an internal electric field is generated in the depletion region.

When the incident light 22 reaches the photoelectric conversion unit 17 in the above situation, electron-hole pairs are generated by photoelectric conversion, and the electrons and holes drift in opposite directions due to an internal electric field. Specifically, electrons drift toward the center of the photoelectric conversion unit 17, and holes drift toward the P-type well. As a result, electrons are accumulated in the photoelectric conversion unit 17, forming signal charges in the photosensitive unit.

Therefore, in the photosensitive cell 1a, signal charges are formed according to the light intensity in the wavelength range of red light in the incident light 22 . Similarly, in the photosensitive cell 1b, signal charges are formed according to the light intensity of the incident light 22 in the wavelength range of red and green light. In the photosensitive unit 1c, signal charges are formed according to the light intensity of the incident light 22 in the wavelength range of red, green and blue light.

However, according to the above structure, the signal charges formed in the photosensitive cells 1b and 1c each include signals of more than one color. Therefore, in order to generate blocky imaging data from the signal charges in the photosensitive cells, it is necessary to derive the color signals (R, G , B). The method of signal processing is explained below.

[signal processing]

FIG. 4 is a block diagram illustrating an internal structure of a signal processing circuit.

The signal processing circuit 6 includes a matrix holding unit 61 , a calculation unit 62 , and a memory control unit 63 .

The matrix holding unit 61 holds a matrix for converting digital signals (Sa, Sb, Sc) generated in the analog front end 5 into color signals (R, G, B).

Calculation unit 62 obtains color signals (R, G, B) by applying matrices to digital signals (Sa, Sb, Sc).

The memory control unit 63 controls access to the work memory 7 and the recording memory 8 .

5A-5B show the matrix and the inverse matrix held in the matrix holding unit 61 .

FIG. 5A shows the relationship between digital signals (Sa, Sb, Sc) and color signals (R, G, B).

In the matrix, W11 and other coefficients represent weighting factors based on the characteristics of the filter layer 20 . For example, when white light enters the photosensitive cell 1c, W11 is 0.333, W12 is 0.333, and W13 is 0.333 in the case where all red light, green light, and blue light pass through without being blocked. In cases where blue light is slightly blocked, some corrections can be made to set the W11 even smaller. Further, in FIG. 5, the numbers of some items of the matrix and the inverse matrix are 0. In practice, however, each term is usually a non-zero number.

The matrix shown in FIG. 5B is the inverse matrix of the matrix held in the matrix holding unit 61 . Therefore, this matrix is obtained by inversely transforming the matrix shown in FIG. 5A. The matrix shown in FIG. 5B is the inverse matrix of the matrix shown in FIG. 5A , which is a matrix held in the matrix holding unit 61 .

[Operation of Memory Control Unit 63 ]

The memory control unit 63 temporarily stores the digital signal from the analog front end 5 in the work memory 7 . When imaging data of one image is stored in the working memory 7 , the memory control unit 63 obtains portions of the imaging data from the working memory 7 and outputs the data to the calculating unit 62 .

The calculation unit 62 applies the matrix held in the matrix holding unit 61 to the input data to obtain color signals (R, G, B).

The memory control unit 63 stores the color signal obtained by the calculation unit 62 in the recording memory 8, and constructs imaging data of one image in the form of the color signal.

In this way, imaging data of one image is recorded in the recording memory 8 .

[Manufacturing method]

Next, a method of manufacturing the filter layer 20 is explained.

An example of a method of manufacturing the filter layer 20 is illustrated in FIGS. 6A-6H.

Figure 6A shows the photosensitive cell after a chemical vapor deposition (CVD) step. In the CVD step, the polysilicon layer 201 is formed by CVD on the entire upper surface of the silicon dioxide layer formed in the light shield forming layer 14 of the photosensitive unit (1a, 1b, 1c). The polysilicon layer 201 is formed to a thickness of 0.5 μm.

Figure 6B shows the photosensitive cell after the photoresist coating step. In the photoresist coating step, a photoresist (PR) 202 is applied to the entire upper surface of the polysilicon layer 201 that has been formed in the CVD step.

Figure 6C shows the photosensitive cell after the exposure/development step. In the exposure/development step, after masking with a mask of a predetermined pattern, the photoresist 202 formed in the photoresist coating step is exposed, and then the exposed portion is removed, and the remaining portion is hardened. Through this process, the photoresist 202 remains only at the portion corresponding to the photosensitive cell 1a, which is not dry-etched in the subsequent dry-etching step.

Figure 6D shows the photosensitive cell after the dry etching step. In the dry etching step, the polysilicon layer 201 after the exposure/development step is etched. The portion of the polysilicon layer 201 where the photoresist 202 does not remain is etched to make the polysilicon layer 201 thinner. The polysilicon layer 201 was etched so that the thickness of the portion corresponding to the photosensitive cells 1b and 1c was 0.3 [mu]m.

When performing dry etching on polysilicon, the film thickness can be controlled with an accuracy of ±30 nm in dry etching.

Figure 6E shows the photosensitive cell after the photoresist coating step. In this step, a photoresist (PR) 203 is applied to the entire upper surface of the polysilicon layer 201, similar to the step shown in FIG. 6B.

Figure 6F shows the photosensitive cell after the exposure/development step. In this step, the photoresist 203 remains only in portions corresponding to the photosensitive cells 1a and 1b, which are not dry-etched in the subsequent dry-etching step.

Figure 6G shows the photosensitive cell after the dry etching step. In the dry etching step, the portion of the polysilicon layer 201 where the photoresist 203 does not remain is etched in order to make the polysilicon layer 201 thinner. The polysilicon layer 201 was etched so that the thickness of the portion corresponding to the photosensitive cell 1c was 0.1 [mu]m.

Figure 6H shows the photosensitive cell after the photoresist removal step. In the photoresist removal step, unnecessary photoresist 203 is removed.

The polysilicon layer 201 formed through the above steps is used as the filter layer 20 shown in FIG. 2 .

As described above, the wavelength range of light transmitted through the filter layer according to the first embodiment is determined by the thickness of the filter layer, not by different colors of pigments or dyes. Therefore, it is unnecessary to control the materials of different colors of pigments or dyes in the manufacturing process, and thus, the production cost can be reduced.

Further, the filter layer can be manufactured using semiconductor manufacturing technology, and it is not necessary to manage the color filter manufacturing process of acrylic resin. As a result, it is possible to utilize manufacturing equipment for semiconductor manufacturing and to simplify the manufacturing process.

In order to reduce light transmission, there are two possible approaches. One is to make the filter layer absorb light, and the other is to make the filter layer reflect light. In the present invention, the former method is used in consideration of material management and occurrence of the glare phenomenon. The glare phenomenon is a phenomenon in which a correct signal overlaps an incorrect signal from reflected light from an adjacent photoelectric conversion unit.

The latter method is realized, for example, by a multilayer structure with two materials of different refractive index arranged alternately. In the former method, as explained above, the filter layer is composed of only one material (polysilicon, amorphous silicon, or silicon). Therefore, the former method is superior to the latter method in terms of material management.

In addition, in the latter method, the light not transmitted through the filter layer can reach the adjacent photoelectric conversion unit by scattering in the photodetector after being reflected by the filter layer. With the former method, the light not transmitted through the filter layer is absorbed by the filter layer, so the glare phenomenon is very unlikely to occur. Thus, the former method is superior to the latter method in terms of the occurrence rate of the glare phenomenon.

In the first embodiment, the main surface of the filter layer 20 facing the light source is covered with a silicon dioxide film having a lower refractive index than the filter layer 20 . Such a silicon dioxide film reduces the refractive index of incident light. Usually, the photodetector is located in a gas such as air, and the light reaches the filter layer after passing through the gas. When the gas and the filter layer 20 have direct contact, the difference between their refractive indices becomes large, so that the reflectance at the boundary becomes high. Therefore, in the first embodiment, the sensitivity is improved by interposing a silicon dioxide film between the gas and the filter layer 20 to reduce the reflectance.

second embodiment

[summary]

In the second embodiment, the filter forming layer 15 is formed under the light shield forming layer 14, and the insulating layer 13 insulating between the filter forming layer 15 and the photodiode layer 12 is formed so as to have a thickness ranging from 1 nm to 150 nm. With the photosensitive cell having such a structure, the reflection of the incident light 22 at the boundary surface between the layers becomes smaller, and therefore, the decrease in sensitivity can be suppressed.

[structure]

Fig. 7 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a second embodiment.

The photosensitive cells according to the second embodiment each include a substrate 11 , a photodiode layer 12 , an insulating layer 13 , a light shield forming layer 14 , and a light filter forming layer 15 . The difference from the photosensitive cell of the first embodiment is that the filter forming layer 15 is formed between the light shield forming layer 14 and the insulating layer 13 . Apart from this difference, the photosensitive unit of the second embodiment is the same as that of the first embodiment. Therefore, only this distinction is explained below.

The filter forming layer 15 includes a filter layer 20 made of polysilicon, and can be formed using common semiconductor manufacturing techniques. Therefore, it is possible to form the light filter forming layer 15 between the photodiode layer 12 and the light shield forming layer 14 .

Since the filter layer 20 is composed of polysilicon as described above, signal charges generated in the photoelectric conversion unit 17 may leak into the filter layer 20 if the filter layer 20 is not insulated from the photodiode layer 12 . Therefore, the insulating layer 13 having a thickness in the range of 1 nm to 150 nm is disposed between the filter layer 20 and the photodiode layer 12 .

The insulating layer 13 having a thickness in the range of 1 nm to 150 nm is a characteristic part of the photosensitive unit of the second embodiment.

The incident light 22 is collected by the microlens 21 and reaches the photoelectric conversion unit 17 through the small hole 19 and the filter layer 20 . In general, when light enters a medium with a higher refractive index from another medium with a lower refractive index, the light is reflected at the boundary surface of the two media in a certain proportion.

The material and refractive index of each layer in the photosensitive unit through which incident light 22 passes are as follows:

Small hole 19: made of silicon dioxide, with a refractive index of 1.45.

Filter layer 20 : made of polysilicon with a refractive index of 4.

Insulating layer 13: made of silicon dioxide, with a refractive index of 1.45.

Photoelectric conversion unit 17 : made of N-type silicon with a refractive index of 4.

Specifically, the incident light 22 passes through the interface S1, that is, the upper surface of the filter layer 20, the interface S2 between the filter layer 20 and the insulating layer 13, and the interface S3 between the insulating layer 13 and the photoelectric conversion unit 17. Scale reflection. Therefore, the light reaching the photoelectric conversion unit 17 decreases due to reflection, and as a result, the sensitivity of the photosensitive unit decreases.

It is known that the refractive index of the boundary surface between insulating layer 13 and photoelectric conversion unit 17 can be reduced by setting the thickness of insulating layer 13 in the range of 1 nm to 150 nm. By setting the thickness of the insulating layer 13 in the range of 1 nm to 150 nm, it is possible to suppress a decrease in the sensitivity of the photosensitive cell.

As explained above, in addition to the effect obtained by the photosensitive unit of the first embodiment, by setting the thickness of the insulating layer 13 in the range of 1 nm to 150 nm so as to reduce the reflection of the incident light 22, the photosensitive unit of the second embodiment It also has the effect of suppressing the reduction in sensitivity of the photosensitive unit.

third embodiment

[summary]

In the third embodiment, the insulating layer 13 between the filter layer 20 and the photodiode layer 12 is removed from the photosensitive cell in order to increase the sensitivity of the photosensitive cell. However, simply removing the insulating layer 13 causes another problem that signal charges generated in the photoelectric conversion unit 17 leak into the filter layer 20 . Therefore, a potential barrier 23 is provided between the photoelectric conversion unit 17 and the filter layer 20 in order to prevent leakage of signal charges. In this way, the reflection at the boundary surface between the insulating layer 13 and the photodiode layer 12 can be completely eliminated.

[structure]

Fig. 8 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a third embodiment.

The photosensitive cells each include a substrate 11 , a photodiode layer 12 , a light shield forming layer 14 , and a light filter forming layer 15 . The difference from the photosensitive cell of the second embodiment is that the insulating layer 13 is not included. Other differences are that a potential barrier 23 is provided between the photoelectric conversion unit 17 and the filter layer 20, and a transfer transistor 24 is provided in each photosensitive unit, as shown in the drawings, for the signal charge generated in the photoelectric conversion unit 17 Toggle between delivery and non-delivery. Except for these differences, the photosensitive unit is the same as that of the second embodiment. Therefore, only the differences are explained below.

A potential barrier 23 is positioned between the photoelectric conversion unit 17 and the filter layer 20 at which a barrier potential is generated, while the potential barrier is formed by doping a P-type impurity in the photoelectric conversion unit 17 . With the potential barrier 23 , the signal charge generated in the photoelectric conversion unit 17 is prevented from leaking to the filter layer 20 . Further, as described above, the barrier 23 is composed of silicon, and its refractive index is substantially the same as that of the filter layer 20 and the photoelectric conversion unit 17 . Accordingly, incident light 22 is hardly reflected at the interfaces (S4, S5) between these layers.

In the transfer transistor 24 , the drain is connected to the photoelectric conversion unit 17 , the source 24S is connected to the horizontal scanning circuit 4 , and the gate 24G is connected to the vertical scanning circuit 3 . Therefore, based on the control signal from the vertical scanning circuit 3 , transfer and non-transfer of signal charges generated in the photoelectric conversion unit 17 can be switched.

Although not shown in the drawings, a transfer transistor 24 is also provided in each of the photosensitive cells 1b and 1c. Furthermore, the transfer transistor 24 is also provided in the photosensitive cell in other embodiments, but is not described in detail because a specific description of the transfer transistor is not necessary in other embodiments.

As described above, in the photosensitive cell, the barrier potential of the barrier 23 prevents signal charges in the photoelectric conversion unit 17 from leaking to the filter layer 20 . Details of the potential barrier 23 will be described below.

9A and 9B show the potentials of the filter layer 20 , the potential barrier 23 , the photoelectric conversion unit 17 and the transfer transistor 24 of the photosensitive unit 20 .

Fig. 9A shows the potential during the non-delivery time.

Barrier 23 is a P-type semiconductor whose potential (barrier potential p2 ) is relatively lower than that of photoelectric conversion unit 17 (photoelectric conversion unit potential p3 ) which is an N-type semiconductor due to the pn junction formed with photoelectric conversion unit 17 . During the non-transfer time, the vertical scanning circuit 3 controls the gate potential p5 of the gate region of the transfer transistor 24 to be low. Therefore, a potential well is formed at the position of the photoelectric conversion unit 17 .

Electrons generated in the photoelectric conversion unit 17 by photoelectric conversion are accumulated in the potential well as signal charges Q1 , and the signal potential p4 drops as the signal charges Q1 accumulate.

Figure 9B shows the potential during delivery.

During the transfer time, the vertical scanning circuit 3 controls so that the gate potential p5 of the gate region of the transfer transistor 24 rises to be the same as the photoelectric conversion unit potential p3. Controlled in this way, the signal charge Q1 migrates toward the source 24S.

As described above, the barrier 23 prevents the signal charge Q1 from leaking to the filter layer 20 by forming the barrier potential p2. In the case where the potential barrier 23 is not provided, when the signal potential p4 falls to be the same as the filter layer potential p1, the signal charge Q1 may leak to the filter layer.

The barrier potential p2 forms a potential well together with the gate potential p5 during the transfer time, so it is desirable to set the barrier potential p2 as high as the gate potential p5 or to make the barrier higher (lower in potential) than the gate potential p5.

As described above, in addition to the same effect as the photosensitive unit of the first embodiment, by removing the insulating layer 13 that insulates the filter layer 20 from the photodiode layer 12, the photosensitive unit of the third embodiment also achieves almost complete The effect of eliminating the reflection at the boundary surface between the insulating layer 13 and the photodiode layer 12 and increasing the sensitivity of the photosensitive unit.

Fourth embodiment

[summary]

The filter layer 20 of the fourth embodiment is formed by doping P-type impurities into the photoelectric conversion unit 17 in the upper portion of the photodiode layer 12 . By this forming method, the manufacturing process can be shortened as compared with the manufacturing process of forming a polysilicon film with a predetermined thickness by CVD and then performing etching so that each film has a predetermined thickness corresponding to each photosensitive unit .

[structure]

Fig. 10 is a cross-sectional view of a photosensitive unit (1a, 1b, 1c) according to a fourth embodiment.

Each photosensitive cell includes a substrate 11 , a photodiode layer 12 , an insulating layer 13 , and a light shield forming layer 14 . The difference from the photosensitive unit of the first embodiment is that the filter layer 20 is formed on the photoelectric conversion unit 17 of the photodiode layer 12 . Other than that, the photosensitive unit of this embodiment is basically the same as that of the first embodiment. The following description deals only with parts different from the first embodiment.

The filter layer 20 is formed by doping P-type impurities into the photoelectric conversion unit 17 . The thicknesses (ta, tb, tc) of the filter layer 20 of the photosensitive unit are set as follows, as in the first embodiment:

ta: 0.5 μm

tb: 0.3μm

tc: 0.1 μm

11A-11C show the relationship between the potential and the distance from the boundary surface S6 between the insulating layer 13 and the photodiode layer 12 in the photosensitive cell.

FIG. 11A shows a photosensitive unit 1a.

The thickness of the filter layer 20 of the photosensitive unit 1a is 0.5 μm, and only light with a wavelength of 580 nm or longer (light in the red wavelength range) of the incident light 22 passes through the filter layer 20 and reaches the photoelectric conversion unit 17 .

In the photoelectric conversion unit 17 , electrons generated by photoelectric conversion drift toward the center of the photoelectric conversion unit 17 and holes drift toward the filter layer 20 due to a potential gradient (internal electric field). In this way, electrons that will become signal charges are accumulated in the photoelectric conversion unit 17 .

Meanwhile, since the filter layer 20 is also made of silicon, electron-hole pairs are generated due to light in the wavelength range of blue light and green light in the incident light 22 . However, the electron-hole pairs do not drift because no potential gradient is formed in the filter layer 20, as shown in FIG. 11A. The electron-hole pairs recombine after a while and then disappear.

Therefore, signal charges in the photoelectric conversion unit 17 of the photosensitive unit 1a are generated based only on light in the red wavelength range.

FIG. 11B shows the photosensitive unit 1b.

The filter layer 20 of the photosensitive unit 1b has a thickness of 0.3 μm, and only light with a wavelength of 490 nm or longer (light within the wavelength range of red and green light) in the incident light 22 passes through the filter layer 20 and reaches the photoelectric conversion unit 17 .

Accordingly, signal charges in the photoelectric conversion unit 17 of the photosensitive unit 1b are generated based on light in the wavelength range of red light and green light.

Figure 11C shows a photosensitive cell 1c.

The thickness of the filter layer 20 of the photosensitive unit 1c is 0.1 μm, and among the incident light 22, only light with a wavelength of 400 nm or longer (light in the wavelength range of red light, green light and blue light) passes through the filter layer 20 and reaches the photoelectric conversion unit 17.

Therefore, signal charges in the photoelectric conversion unit 17 of the photosensitive unit 1c are generated based on light in the wavelength ranges of red light, green light and blue light.

[Manufacturing method]

Next, a method of manufacturing the filter layer 20 according to the present embodiment is explained.

12A-12G illustrate another example of a method of manufacturing the filter layer 20. As shown in FIG.

The filter layer 20 is formed by implanting P-type impurity ions after the photoelectric conversion unit 17 is formed.

Figure 12A shows the photosensitive cell after the ion implantation step.

In the ion implantation step, N-type impurities such as phosphorus and arsenic or P-type impurities such as boron are ionized, accelerated by an electric field, and implanted into a silicon substrate or the like. Areas not subjected to ion implantation are covered and protected by photoresist 401 .

In the embodiment shown in the drawings, the photoelectric conversion unit 17 is formed by doping the P-type well 16 with N-type impurities.

Figure 12B shows the photosensitive cell after the ion implantation step.

In the ion implantation step, a P-type region 402 having a thickness of 0.1 μm is formed by ion implantation of P-type impurities into the photoelectric conversion unit 17 using the photoresist 401 used when forming the photoelectric conversion unit 17 . Thus, the filter layer of the photosensitive cell 1c is formed.

During ion implantation, the variation of film thickness is precisely controlled according to the type of implanted ions. For example, when boron is selected as the P-type impurity, the accuracy of ±3nm can be controlled.

Figure 12C shows the photosensitive cell after the photoresist coating step.

In the photoresist coating step, photoresist 403 is applied to the entire upper portion of the substrate.

Then, in the exposure/development step, the photoresist 403 applied to the portion where ion implantation is to be performed is removed.

Figure 12D shows the photosensitive cell after the ion implantation step.

By doping P-type impurities in the photoelectric conversion unit 17, a P-type region 404 with a thickness of 0.3 μm is formed. Thus, the filter layer of the photosensitive cell 1b was formed. The photosensitive unit 1c is covered and protected by a photoresist, so no ion implantation is performed on the photosensitive unit 1c.

Figure 12E shows the photosensitive cell after the photoresist coating step. In the photoresist coating step, photoresist 405 is applied to the entire upper portion of the substrate.

Then, in the exposure/development step, the photoresist 405 applied to the portion where ion implantation is to be performed is removed.

Figure 12F shows the photosensitive cell after the ion implantation step.

By doping P-type impurities in the photoelectric conversion unit 17, a P-type region 406 having a thickness of 0.5 μm is formed. Thus, the filter layer of the photosensitive cell 1a was formed. The photosensitive units 1b and 1c are covered and protected by photoresist, so no ion implantation is performed on the photosensitive units 1b and 1c.

Figure 12G shows the photosensitive cell after the photoresist removal step. In the photoresist removal step, the photoresist 405 is removed since it is no longer necessary. Thus, the filter layer 20 is formed.

As explained above, in addition to realizing the effect of the photosensitive unit of the first embodiment, the photosensitive unit of the third embodiment also realizes the effect of shortening the manufacturing process to a large extent because the filter layer 20 is formed by using the photoelectric conversion unit 17 using The same photoresist to form.

Variations

Although the imaging system of the present invention has been explained based on the preferred embodiments, the present invention is not limited to the above-described embodiments. Other possible forms of the invention include variant embodiments as described below.

(1) In the preferred embodiment, a MOS image sensor is taken as an example for description. However, the present invention is also applicable to CCD image sensors.

(2) The material used for the filter layer 20 is not limited to polysilicon, and may be any material as long as light in different wavelength ranges passes through the material according to the film thickness. For example, amorphous silicon may be used for the filter layer 20 .

(3) The filter layer 20 is made of polysilicon only. However, the present invention is not limited to the embodiment if each filter layer can absorb light having a wavelength greater than a predetermined wavelength more than light having a wavelength smaller than the wavelength. For example, the filter layer may be in the form of a multilayer laminate composed of different materials.

(4) The filter layer 20 corresponding to the photosensitive unit 1c also has a function of cutting off light in the ultraviolet region. Therefore, in the case that the imaging system includes other devices for cutting off ultraviolet light, the filter layer 20 corresponding to the photosensitive unit 1c does not necessarily have to be provided. In an ordinary imaging system, the incident light gathered by the lens reaches the photodetector 1 after passing through the optical low-pass filter and the infrared cut-off filter. In this imaging system, if a UV cut filter is also provided, there is no need to cut UV light at the photosensitive unit 1 c of the photodetector 1 .

Fig. 13 is a cross-sectional view of the photosensitive unit (1a, 1b, 1c) of the embodiment in which the imaging system is provided with an ultraviolet cut filter.

As shown in Figure 13, the thickness of the filter layer 20 corresponding to the photosensitive unit 1a is 0.5 μm, the thickness of the filter layer 20 corresponding to the photosensitive unit 1b is 0.3 μm, and the part corresponding to the photosensitive unit 1c is not provided with a filter layer .

(5) In the first to third embodiments, the filter layer 20 is formed integrally on the photosensitive unit. Therefore, it is possible to form a potential gradient in the filter layer 20 by setting one end of the filter layer 20 to be connected to the ground potential (0V) as a whole, and the other end to be connected to the source potential (for example, 3V). When light passes through the filter layer 20, electrons of electron-hole pairs generated in the filter layer 20 are absorbed to the power supply side, and holes are absorbed to the ground side.

(6) In order to apply the present invention to an ordinary digital still camera or digital video camera, the thickness of the filter layer 20 in the photosensitive unit is set to be ta=0.5 μm, tb=0.3 μm, and tc=0.1 μm. However, in the case where the present invention is used for other purposes, the thickness is not limited to the above-mentioned numerical value.

(7) Strictly speaking, as shown in FIG. 3 , a small amount of light having a wavelength smaller than the cutoff wavelength of 580 nm passes through the filter layer having a thickness of 0.5 μm. Specifically, in addition to light in the red wavelength range, light in the blue or green wavelength range passes through the filter layer of the photosensitive unit 1a to some extent. However, the signal charge based on light having a wavelength smaller than the cutoff wavelength is as small as zero and does not contribute much in forming the signal charge. Therefore, in order to simplify the description, only the light in the red wavelength range passes through the filter layer of the photosensitive unit 1 a in this specification. Similarly, the small amount of light having a wavelength smaller than the cut-off wavelength that has passed through the filter layers of the photosensitive units 1b and 1c is not considered in the specification.

The matrices and inverse matrices shown in the embodiment include entries of 0 for the reasons described above. However, if the transmission of light having a wavelength smaller than the cutoff wavelength is considered, each term may have a value other than 0.

(8) In the embodiment, the filter layer 20 is formed integrally on the photosensitive units 1a, 1b, and 1c. However, the filter layer 20 may also be formed independently.

(9) In the embodiment, silicon dioxide is used as a material for the antireflection layer on the main surface of the filter layer 20 facing the light source. However, the material of the anti-reflection layer is not limited to silicon dioxide if the refractive index of the material is smaller than that of the filter layer 20 . For example, the anti-reflection layer may be composed of silicon nitride or silicon oxynitride (SiON).

Although the present invention has been fully described through the embodiments with reference to the accompanying drawings, it is to be noted that various changes and modifications will be apparent to those skilled in the art. Therefore, unless otherwise these changes and modifications depart from the scope of the present invention, they should be construed as included therein.

Claims (21)

1. a photodetector comprises the Semiconductor substrate with a plurality of photosensitive units, and each photosensitive unit comprises:

Filter layer is used for the light in the transmission photosensitive unit predetermined wavelength range; With

Photoelectric conversion unit is used for producing signal charge according to the transmission light intensity,

Wherein

The thickness of each filter layer is corresponding to each photosensitive unit presetted wavelength scope, and

Each filter layer in all photosensitive units is all only made by a kind of material.

2. photodetector according to claim 1 is characterized in that,

Filter layer is made by a kind of material, the transmissivity of this material when the transmissivity of optical wavelength during less than cut-off wavelength is lower than optical wavelength and is equal to or greater than cut-off wavelength, and described cut-off wavelength is determined by the thickness of this material.

3. photodetector according to claim 2 is characterized in that filter layer is made by the material that absorbing light reduces by transmissivity.

4. photodetector according to claim 2 is characterized in that filter layer is made of one of polysilicon, amorphous silicon and silicon.

5. photodetector according to claim 3 is characterized in that,

Substrate is made of silicon,

Photoelectric conversion unit form by doped N-type impurity in substrate and

Filter layer forms by doping p type impurity in photoelectric conversion unit.

6. photodetector according to claim 2 is characterized in that, filter layer becomes big material by its cut-off wavelength along with the material thickening and makes.

7. photodetector according to claim 2 is characterized in that,

The thickness of each filter layer is one of first thickness, second thickness and the 3rd thickness,

The cut-off wavelength of first thickness between red light wavelength scope and green wavelength,

The cut-off wavelength of second thickness between green wavelength and blue light wavelength scope, and

The cut-off wavelength of the 3rd thickness is between blue light wavelength scope and UV wavelength range.

8. photodetector according to claim 1 is characterized in that filter layer is made of one of polysilicon, amorphous silicon and silicon.

9. photodetector according to claim 1 is characterized in that,

Substrate is made of silicon,

Photoelectric conversion unit form by doped N-type impurity in substrate and

Filter layer forms by doping p type impurity in photoelectric conversion unit.

10. photodetector according to claim 1 is characterized in that,

Photosensitive unit further comprises anti-reflecting layer, and it is made less than the filter layer refractive index materials by refractive index, and anti-reflecting layer is arranged on filter layer on the first type surface of light source.

11. photodetector according to claim 10 is characterized in that,

Filter layer constitute by one of polysilicon, amorphous silicon and silicon and

Anti-reflecting layer is made of one of silicon nitride, silicon dioxide and silicon oxynitride.

12. photodetector according to claim 1 is characterized in that,

Photosensitive unit further comprises the light shield cambium layer, and it has light shield layer and aperture in the part corresponding to photoelectric conversion unit, light shield layer hinder by the light beyond the aperture and

Filter layer is positioned between light shield cambium layer and the photoelectric conversion unit.

13. photodetector according to claim 12 is characterized in that,

Photosensitive unit further comprises silicon dioxide layer, and its thickness, is arranged between filter layer and the photoelectric conversion unit in the scope of 150nm at 1nm.

14. photodetector according to claim 12 is characterized in that,

Photosensitive unit further comprises:

Grid when signal charge is not carried, can be operated the gate regions that is used between photoelectric conversion unit and conveying target and produce grid potential, and grid potential is lower than the electromotive force at photoelectric conversion unit place; With

Potential barrier is arranged between filter layer and the photoelectric conversion unit, produces the potential barrier electromotive force that is lower than grid potential at the potential barrier place.

15. a photodetector comprises the Semiconductor substrate with a plurality of photosensitive units, each photosensitive unit comprises:

Photoelectric conversion unit can be operated and is used for producing signal charge according to light intensity, wherein

Each all is provided with filter layer in the position that sees through light to photoelectric conversion unit some photosensitive units, the thickness of each filter layer is corresponding to each photosensitive unit presetted wavelength scope, each filter layer in all more described photosensitive units all only make by a kind of material and

Remaining photosensitive unit is not provided with filter layer.

16. photodetector according to claim 15 is characterized in that,

Filter layer is made by a kind of material, the transmissivity of this material when the transmissivity of optical wavelength during less than cut-off wavelength is lower than optical wavelength and is equal to or greater than cut-off wavelength, and described cut-off wavelength is determined by the thickness of material.

17. photodetector according to claim 16 is characterized in that,

The thickness of each filter layer is one of first thickness and second thickness,

The cut-off wavelength of first thickness between red light wavelength scope and green wavelength and

The cut-off wavelength of second thickness is between green wavelength and blue light wavelength scope.

18. signal processing apparatus, obtain color signal based on first source signal and second source signal in the multiple source signals that produces by photodetector as claimed in claim 1, first source signal is corresponding to the light intensity in first wave-length coverage, second source signal is corresponding to the light intensity in second wave-length coverage that comprises first wave-length coverage, color signal is corresponding to the light intensity in second wave-length coverage of getting rid of first wave-length coverage, and signal processing apparatus comprises:

Holding unit keeps respectively the weight coefficient corresponding to first source signal and second source signal; With

Computing unit can be operated and is used for obtaining difference between the multiplied result then, thereby obtaining color signal by first source signal and second source signal are multiplied by weight coefficient respectively.

19. signal processing apparatus, based on first source signal in the multiple source signals that produces by photodetector as claimed in claim 1, second source signal and the 3rd source signal obtain the danger signal corresponding to light intensity in the red light wavelength scope, green corresponding to light intensity in the green wavelength, with blue signal corresponding to light intensity in the blue light wavelength scope, first source signal is corresponding to the light intensity of wave-length coverage greater than the light between red light wavelength scope and the green wavelength, second source signal is corresponding to the light intensity of wave-length coverage greater than the light between green wavelength and the blue light wavelength scope, the 3rd source signal is corresponding to the light intensity of wave-length coverage greater than the light between blue light wavelength scope and the UV wavelength range, and signal processing apparatus comprises:

Holding unit keeps one group of first source signal, and second source signal and the 3rd source signal are converted to one group of danger signal, the transition matrix of green and blue signal; With

Computing unit can be operated and is used for that transition matrix is applied to this and organizes first source signal, second source signal and the 3rd source signal.

20. method of making photodetector, wherein photodetector comprises the Semiconductor substrate with a plurality of photosensitive units, each photosensitive unit is provided with the filter layer of light in the transmission photosensitive unit predetermined wavelength range, and photoelectric conversion unit, can operate and be used for producing signal charge according to the transmission light intensity, this method comprises:

On each photosensitive unit, form photoelectric conversion unit;

Form the layer of same thickness above each photoelectric conversion unit, this layer is made by a kind of material, and the wave-length coverage of the transmitted light of this material is determined based on thickness; With

In all photosensitive units, form each filter layer of only making by the described layer of etching, be each photosensitive unit preset thickness so that the layer that obtains has by a kind of material.

21. method of making photodetector, wherein photodetector comprises the Semiconductor substrate with a plurality of photosensitive units, each photosensitive unit is provided with the filter layer of light in the transmission photosensitive unit predetermined wavelength range, and photoelectric conversion unit, can operate and be used for producing signal charge according to the transmission light intensity, this method comprises:

Doped N-type impurity forms photoelectric conversion unit in each photosensitive unit; With

In all photosensitive units, form each filter layer of only making by doping p type impurity in photoelectric conversion unit, doped portion is had be each photosensitive unit preset thickness by a kind of material.

Images